Politicians Are Trying to Change What Chatbots Say About Them

Political campaigns have discovered a new battleground: the training data and outputs of AI chatbots. When voters ask ChatGPT, Perplexy, or Claude about a candidate, campaigns want to make sure the answer isn't unflattering, and they're starting to pressure AI companies directly to change what their models say.

This isn't about deepfakes or misinformation campaigns. It's about something more subtle: politicians trying to shape the "official" AI-generated summary of their record and positions. The worry for campaigns is that a chatbot might surface an old controversy, omit key achievements, or present an opponent's framing of their record. And unlike a bad search result you can push down with SEO, there's no clear playbook for influencing what an LLM says about you.

The bigger issue here is accountability. Search engines have established processes for handling inaccurate information and legal requests. AI companies are still figuring this out. Some campaigns are reportedly reaching out directly to companies like OpenAI and Anthropic, but there's no transparency about who gets heard and what changes get made. That creates a risk that well-funded campaigns get preferential treatment in shaping AI outputs.

This also puts AI companies in an uncomfortable position. They're already struggling with accuracy and bias in political content. Now they have to decide whether to treat candidate information like any other factual query or create special review processes for political figures. Either choice has consequences for how voters get information.

For anyone building with AI, this is a reminder that LLMs are not neutral reference sources. They reflect choices about training data, system prompts, and moderation policies. When you're using AI for research or decision support, cross-reference claims, especially on contested topics like politics, business reputation, or anything where someone has an incentive to shape the narrative.

What this means for you: if you're using AI tools for research on people, companies, or controversial topics, always verify key facts with primary sources. Try this prompt when you need balanced information: "Give me multiple perspectives on [topic], including criticisms and counterarguments, and cite specific sources for each claim." It won't eliminate bias, but it surfaces what the model is drawing from and makes gaps more obvious.